Research On The OpenFlow Network Qos Guarantee Subsystem

The growing emergence of a variety of network applications have stringent Quality of Service(QoS) requirements which cannot be always met by the best-effort Internet.To provide QoS, several QoS frameworks have been explored over last two decades,but none of them has been truly successful and globally implemented. OpenFlow is a new network switch model whose management and control is separated, it can provide a flexible flow definition, and its centralized network control can achieve the latest state of the global network. In addition, OpenFlow has the ability to manage network devices instantaneously to adapt to sewn end to end network behavior, therefore, any QoS mechanism or framework on OpenFlow will be able to support the deployment of end to end QoS. At present, there are few research on QoS architecture based on OpenFlow network at home and abroad, and these studies mainly use the resource reservation,priority queues and QoS routing to provide QoS guarantees for applications. However,the former two would have a greater impact on flows without QoS requirements, while QoS route does not really combine with the current state of the network, so the calculated routes are lacking in precision.Considering above problems, this dissertation proposes a new QoS guarantee subsystem according to the features of OpenFlow’s global control and support for statistics collection, the QoS guarantee subsystem mainly collectes the real-time network QoS parameter information, while these information is referred by QoS routing to select routes satisfying applications’ QoS requirement. The subsystem consists of five major modules: route calculation, network measurement, access control, flow management, topology management. The new QoS guarantee subsystem presented in this dissertation is implemented in the FloodLight controller development platform.This dissertation realizes the path calculation and network measurement, and modifies the topology manager module. Path calculation is used to calculate the path to meet user QoS requirements, based on the study after a variety of QoS routing algorithm, this dissertation employees NR_DCLC optimization algorithm to calculate the path for QoS applications. Network measurement is used for collecting current network status information, it is divided into three sub-modules, including the link measurement module, packet loss measurement module and delay measurement module. Pathcalculation module utilizes information collected by measurement system to calculate an optimized QoS routing, when QoS network measurement examines current resource is unable to meet the requirements of the application, it will notify the path calculation module to re-select a new path. This QoS guarantee subsystem is simulated and verified on Mininet simulation platform. Simulation results show that the network measurement can correctly obtain network status information, QoS routing approach can improve resource utilization and can achieve dynamic re-routing in the case of network congestion.Although the new QoS guarantee subsystem presented in this dissertation is implemented on simulation environment, the design and implementation of the function and the code can be applied to real network environment, only if each switch supports OpenFlow protocol. So the working has great significance for theoretical research and practical deployment.